Single-Molecule Atomic Force Microscopy Force

ARTICLE pubs.acs.org/biochemistry

Single-Molecule Atomic Force Microscopy Force Spectroscopy Study of Aβ-40 Interactions Bo-Hyun Kim,† Nicholas Y. Palermo,‡ Sandor Lovas,‡ Tatiana Zaikova,§ John F. W. Keana,§ and Yuri L. Lyubchenko*,† †

Department of Pharmaceutical Sciences, University of Nebraska Medical Center, 986025 Nebraska Medical Center, Omaha, Nebraska 68198, United States ‡ Department of Biomedical Science, Creighton University, Omaha, Nebraska 68178, United States § Department of Chemistry, 1253, University of Oregon, Eugene, Oregon 97403-1253, United States

bS Supporting Information ABSTRACT: Misfolding and aggregation of amyloid β-40 (Aβ-40) peptide play key roles in the development of Alzheimer’s disease (AD). However, very little is known about the molecular mechanisms underlying these molecular processes. We developed a novel experimental approach that can directly probe aggregation-prone states of proteins and their interactions. In this approach, the proteins are anchored to the surface of the atomic force microscopy substrate (mica) and the probe, and the interaction between anchored molecules is measured in the approach
retraction cycles. We used dynamic force spectroscopy (DFS) to measure the stability of transiently formed dimers. One of the major findings from DFS analysis of R-synuclein (R-Syn) is that dimeric complexes formed by misfolded R-Syn protein are very stable and dissociate over a range of seconds. This differs markedly from the dynamics of monomers, which occurs on a microsecond to nanosecond time scale. Here we applied the same approach to quantitatively characterize interactions of Aβ-40 peptides over a broad range of pH values. These studies showed that misfolded dimers are characterized by lifetimes in the range of seconds. This value depends on pH and varies between 2.7 s for pH 2.7 and 0.1 s for pH 7, indicating that the aggregation properties of Aβ-40 are modulated by the environmental conditions. The analysis of the contour lengths revealed the existence of various pathways for dimer dissociation, suggesting that dimers with different conformations are formed. These structural variations result in different aggregation pathways, leading to different types of oligomers and higher-order aggregates, including fibrils.

P

rotein misfolding and self-assembly into various morphologic aggregates are widespread phenomena in the development of various neurodegenerative disorders such as Alzheimer’s and Parkinson’s diseases.1,2 The aggregation properties of amyloid β (Aβ) peptide, a key protein for Alzheimer’s disease, have been studied using various methods.3
7 The current model for amyloidogenic peptides dissects the aggregation kinetics into two main phases. Historically, self-assembly kinetics were observed initially in the earlier work of Hofrichter et al.,8 in which the gelation phenomenon of purified deoxyhemoglobin was investigated. The authors proposed a model that dissected the growth kinetics of fibrils into two phases. The first phase is the nucleation process. During this phase, a critical oligomer of a particular size (nucleus) is formed. The nucleus undergoes a thermodynamically favorable elongation process in which monomers are added via consecutive steps. This model, applied to the aggregation of deoxyhemoglobin, suggested the size of the nucleus to be as large as 30 monomeric units.8 Jarrett and Lansbury applied this model to analyze the aggregation of amyloids that followed a similar kinetic profile.9 It was recently shown that growth of amyloid plaques in vivo follows the same r 2011 American Chemical Society

model.10 Both in vivo and in vitro studies indicate a considerably long lag period during which stable nuclei form. This period is considered a key step in the process of amyloid growth. However, a number of important questions arise. What are these nuclei? How large are they? How long do they live? Answering these questions is important, as soluble oligomeric assemblies of Aβ rather than fibrils are neurotoxic11 and the dimer is the smallest synaptotoxic species.12,13 Progress has been made recently in understanding the early stages of aggregation and misfolding of proteins. The ionization mass spectrometry (ISI-MS) method was developed to characterize oligomeric species of Aβ peptides.3 This ISI-MS using study shows that oligomerization of Aβ-40 and that of Aβ-42 follow two different aggregation pathways with tetramers and hexamers as potential nuclei. Importantly, in both cases, dimers are the building blocks. Spectroscopic analysis of covalently cross-linked Aβ-42 oligomers4,14 showed that changes in the Received: January 28, 2011 Revised: May 6, 2011 Published: May 10, 2011 5154

dx.doi.org/10.1021/bi200147a | Biochemistry 2011, 50, 5154–5162

Biochemistry structure of Aβ-42 occur rapidly in the transition from monomer to dimer and then proceed gradually for higher-order oligomers. A combined approach utilizing several methods to study entire aggregation kinetics has been recently proposed.15 The proposed kinetic model suggests that the β-LGA monomers are converted into dimers and tetramers in the early stages of aggregation, and these species, primarily tetramers, constitute a reservoir of intermediates used for the later stages of the aggregation process. A common feature exists in the studies performed on two different types of amyloidogenic proteins, the conversion of monomers into dimers, followed by their assembly into tetramers. No trimers are detected in either study. These findings suggest that dimers may contain a property that makes their accumulation preferable compared to the use of monomers as building materials for the growth of oligomers. Recently, a novel approach to characterizing misfolded states and dimeric forms of the amyloidogenic proteins was proposed.16,17 In this technique, the interaction between the proteins is measured by AFM force spectroscopy with significant rupture forces, suggesting that misfolded proteins form dimeric states. This approach was tested initially with the use of three different proteins, R-synuclein, lysozyme, and amyloid β peptide.16 Further improvement of the protein immobilization methodology made it possible to provide the characterization of misfolded dimers at the single-molecule level.18 The dynamic force spectroscopy (DFS) approach was used to characterize interactions of R-synuclein (R-Syn).19,20 One of the major findings from DFS analysis is that dimeric complexes formed by misfolded R-Syn protein are very stable and dissociate over a range of seconds. This differs markedly from the dynamics of monomers, which occurs on a microsecond to nanosecond time scale. This finding, along with other recent publications,3,15 suggests that dimerization is the key step in the self-assembly process. Here, we have applied the AFM force spectroscopy methodology to test whether the interaction of the Aβ-40 peptide follows the same model. We used DFS to measure the stability of transiently formed Aβ-40 dimers. The aggregation properties of Aβ-40 depend on structural properties of Aβ-40 that are modulated by the environmental conditions.21
23 Therefore, we performed DFS analysis over a range of pH values. These experiments showed that like R-Syn protein, Aβ-40 forms dimers with lifetimes in the range of seconds. The analysis of the contour lengths revealed variations in the interactions of Aβ-40. This suggests that dimers with different conformations are formed, which we propose are capable of leading to diverse aggregation pathways of Aβ-40.

’ MATERIALS AND METHODS Synthesis of Cysteine-Modified Aβ-40. The procedure is described in detail in the Supporting Information. Briefly, the cysteine-modified N-terminus of Aβ-40 (Cys-Aβ-40) was synthesized on a CEM Liberty microwave peptide synthesizer using a Val-HMPB Chemmatrix resin. We cleaved the synthesized peptide from the resin by stirring the peptide resin with a TFA/ thioanisole/phenol/H2O/dimethyl sulfide/ethanedithiol/triisopropylsilane mixture. After cleavage, the peptide was dissolved in 100 μL of TFA, which was then diluted with 100 mL of H2O and injected into a Vydac C4 semipreparative RP-HPLC column for purification (see the details in the Supporting Information). The purified peptide was characterized by ESI mass spectrometry

ARTICLE

and sodium dodecyl sulfate
polyacrylamide gel electrophoresis (Figure S1 of the Supporting Information). Silatrane-Derivatized Tetrahedrally Shaped (T-silatrane) Molecule. The nanoscale tetrahedrally shaped tripodal silatrane incorporating a chemically reactive terminal maleimide group (Figure S2 of the Supporting Information) has been synthesized as described in the Supporting Information and used without extra purification. Tip and Mica Surface Modification. The general process for the cleaning and modification of the tip and mica surfaces was similar to the previously reported protocol.19,20 Briefly, tips (MLCT, Veeco, Santa Barbara, CA) were cleaned with ethylene alcohol followed by UV treatment for 30 min and modified with maleimide polyethylene glycol silatrane (MAS).18 The freshly cleaved mica surface was modified with 1-(3-aminopropyl)silatrane (APS), and then N-hydroxysuccinimide-polyethylene glycol-maleimide (NHS-PEG-MAL) (MW of 3400 g/mol, Laysan Bio Inc., Arab, AL) was coupled to the amine group of APS in DMSO in a dry chamber for 3 h. Both functionalization protocols yielded surfaces terminated with maleimide headgroups capable of covalent bonding with Aβ-40 at the N-terminal cysteine. To accomplish this step, the functionalized tip and mica were incubated with an ∼20 nM solution of Aβ-40 diluted in HEPES buffer (pH 7) for 1 h and then gently rinsed with the dilution buffer and NaCO3/NaHCO2 (pH 10) buffer. Prepared tips and mica were kept in HEPES buffer until needed. Modification of the Si-Wafer with Tetrahedrally Shaped Tripodal Silatrane [T-shaped (see Figure S2 of the Supporting Information)]. N-Type Si-wafer (NOVA electronic materials, Flower Mound, TX) was cleaned using the standard cleaning process.24 Before the standard cleaning process, the wafer was cleaned with piranha solution (3:1 sulfuric acid/hydrogen peroxide mixture) for 30 min and then rinsed with DI water several times. Precleaned Si-wafer was immersed in SC1 cleaner (5:1:1 water/ammonium hydroxide/hydrogen peroxide mixture) and then SC2 cleaner (6:1:1 water/hydrogen chloride/hydrogen peroxide mixture) at ∼70 °C for 15 min for each step. After each cleaning step, wafer was washed with DI water. The cleaned Si-wafer surface was modified with T-shaped silatrane. The molecule was dissolved at a concentration of 1 mg/mL in DMSO as a stock solution and then diluted in a DMSO/H2O mixture (80:20) before use. The cleaned wafer (1 cm  1 cm) was covered with 30 μL of a diluted T-shaped molecule (167 μM) and was kept in a closed chamber overnight. After being modified, the wafer was rinsed several times using DMSO and then DI water. On the modified Si-wafer surface, Aβ protein was immobilized through the process described above. The working buffer solutions were prepared at five different pHs: pH 2.7 (20 mM glycine-HCl), pH 3.7 (10 mM sodium acetate), pH 5 (10 mM sodium acetate), pH 7 (20 mM HEPES), and pH 9.8 (100 mM sodium carbonate and sodium bicarbonate). All buffers were adjusted to an ionic strength of 150 mM with NaCl. Single-Molecule Force Spectroscopy. The force
distance curve (FDC) of Aβ was measured at four different pH values (2.7, 3.7, 5, and 7) with a MFP3D AFM instrument (Asylum Research, Santa Barbara, CA) at room temperature. The spring constant of tips (MLCT, Veeco) was 30
60 pN/nm, which was calibrated with the thermal noise analysis method. We applied a 100 pN force (trigger) at the contact. At high retracting velocities (g1 μm/s), the tip was kept at the surface for 0.3 s (dwell time). The retracting velocity varied from 50 nm/s to 3 μm/s with 5155

dx.doi.org/10.1021/bi200147a |Biochemistry 2011, 50, 5154–5162

Biochemistry

ARTICLE

Figure 1. (A) Schematic experimental system of SMFS. The AFM tip and mica surface were functionalized with MAS and Aβ-40 and with APS, PEG, and Aβ-40, respectively. In the schematic, Aβs are shown as β-hairpin structures in dimeric form. (B) Representative force
distance curve measured at pH 5 with an apparent loading rate (ALR) of 4280 pN/s. (C) Histograms of the rupture force distribution measured at pH 5 at ALRs of 4000, 12690, 46210, and 65160 pN/s. The most probable rupture forces (Fr) for each loading rate were ∼88, ∼102, ∼143, and ∼177 pN, respectively, which were calculated by Gaussian fitting.

seven or eight discrete steps corresponding to an apparent loading rate of 1000
200000 pN/s. FDCs were collected more than 2000 times per each retracting velocity on five to seven different batches. The selected FDCs were analyzed using the wormlike chain (WLC) model with Igor Pro version 6.01. Data Analysis. The contour length analysis was performed with the WLC model for flexible linkers25 as described in previous publications.18,19,26,27 These papers also describe specifics for DFS methods used in this paper. While the individual FDCs were fitted to the WLC model [Fp = 1/4kBT(1
x/L)
2
1/4 þ x/L, where F is rupture force, p is persistence length, kBT is thermal energy, and L is contour length], the persistence length was allowed to be varied for the best fitting curve and evaluated as a variable parameter along with contour length. The WLC fitting procedure as a part of the software was provided by the MFP 3D manufacturer (Asylum Research). An example of such a fit is shown in Figure S6A of the Supporting Information. The persistence length was an adjustable parameter with a mean value 0.16 ( 0.1 nm in the distribution histogram (Figure S6B of the Supporting Information). On the basis of the parameters of the DFS result, the conceptual simplified energy landscape was constructed on the reaction coordinate.28

’ RESULTS Experimental Design. Figure 1A provides the schematics for our experimental approach. We used a variant of the Aβ-40 peptide with cysteine at the N-terminus to specifically immobilize proteins at the tip and the mica surface functionalized by maleimide using appropriate linkers.18
20,27 This immobilization

approach was selected because the N-terminus of Aβ-40 is not involved in fibril formation.6,29,30 Aβ-40 was covalently attached to the maleimide-terminated surface at its N-terminal cysteine, and the interaction between the protein molecules was measured by multiple approach
retraction cycles. The cysteine sulfhydryl group was attached to the mica surface via a polyethylene glycol linker (PEG, 19
26 nm long). A shorter PEG spacer of maleimide silatrane (MAS) provided immobilization of Aβ-40 to the AFM tip using a one-step modification procedure.18 A freshly prepared ∼20 nM working solution of peptide was used in each experiment, and 0.25 mM TCEP was added to minimize the dimerization of the peptide via S
S bond formation. Note that the concentration of the protein was more than 3 orders of magnitude less than the protein concentration used in the aggregation experiments in vitro. Under these conditions, we obtained a sparse distribution of the peptide on the surface minimizing the double-rupture events.26 Interactions between Aβ-40 at pH 5. A typical force
distance curve for the rupture events of the experiments performed at pH 5 is shown in Figure 1B. pH 5 corresponds to conditions at which Aβ-40 is mostly neutral (pI 5.4). The first large peak of the FDC corresponds to the short-range nonspecific interactions typically appearing in AFM force measurement experiments.20,27 The second peak is separated from the initial peak by a specific distance, similar to prior studies.19,20,26 This distance is defined primarily by the length of the flexible linkers and unstructured segments of the peptide. The peak shown in Figure 1B corresponds to a contour length Lc of ∼39 nm, which is in the range of expected values (Figure 1A and Table S1 of the Supporting Information). The force curves similar to the one in 5156

dx.doi.org/10.1021/bi200147a |Biochemistry 2011, 50, 5154–5162

Biochemistry

ARTICLE

Figure 2. Results of dynamic force spectroscopy analysis at pH 5. (A) Dependence of Fr on the logarithm of ALR. The solid lines represent the best fits of data points in two regimes by Bell’s model. (B) Profile of the energy landscape calculated from the DFS plot above. The parameters are listed in Table 1.

this figure were acquired by multiple probings over various positions of the tip over the surface. The DFS analysis required the system pulling with various rates. Therefore, the probing experiments were performed over a broad range of loading rates (0.1
100 nN/s). To generate the DFS spectrum, we collected more than 25000 force curves. The mean yield of the rupture events was 6.5%, and they all were analyzed. In the control experiments without peptides, nonspecific rupture events appeared at short contour lengths (